Study on the Spatial Correlation between Artificial Intelligence Patent Growth and Regional Innovation Capabilities Based on the Spatial Durbin Model and Geographically Weighted Regression Model

Abstract

This study aims to deeply analyze the complex spatial linkage mechanism between the growth of artificial intelligence patents in China and regional innovation capacity. Based on panel data from Chinese provinces and municipalities from 2019 to 2023, this research comprehensively employs two spatial econometric methods: the Spatial Durbin Model (SDM) and the Geographically Weighted Regression Model (GWR). The study first reveals significant spatial dependence in regional innovation factors through the Moran's I index. Subsequently, SDM results demonstrate that R&D investment exerts a significant positive direct effect on local patent output, yet its spatial spillover effects exhibit substantial fluctuations—shifting from a negative "siphon effect" in 2019 to positive spillovers in subsequent years, reflecting dynamic changes in interregional competition and cooperation. GWR model analysis further confirms that the impact of university and scientific personnel concentration on R&D intensity exhibits significant spatial heterogeneity, highlighting uneven regional innovation capacity development. The ultimate objective is to provide theoretical support and decision-making basis for governments to formulate scientifically sound regional innovation policies and optimize artificial intelligence industry distribution.